Issue link: https://iconnect007.uberflip.com/i/255022
February 2014 • The PCB Magazine 65 ARE BORINGLy PREDICTABLE OPERATIONS POSSIBLE? continues Gray mcQuarrie is president of Grayrock & associates, a team of experts dedicated to building collaborative team environments that make companies maximally effec- tive. to read past columns, or to contact mcQuarrie, click here. ture with dynamic tables (tables that grow and shrink based on conditions as a model run pro- ceeds). This allows for an accounting of a high level of detail, an ability to make changes easily and quickly, and transparency to see and vali- date what is going on as it is happening. For example, as jobs come into layup, they get bro- ken down into individual panels and become part of a bin of material or what I call the layup queue. Table 1 shows the panels selected for the sec- ond press load in the layup queue. This table shows there are a total of 34 panels in the layup queue consisting of a certain number of prod- uct C and product D panels. According to the "NumOfLams" column, all product in the lay- up queue at this point in time are at their first pressing. The last column on the right shows the press load ID column. The next column to the left shows the Book ID number. Now earlier in the run, the first press load went to press 1. Table 2 shows the lamina- tion press load status with the first press load in Press A, which is a five opening press (as seen in the "NumofOpenings" column). Since product C, for its first lamination pressing, has 4 panels per book, there should be 5 books or 20 panels for this first press load. Table 3 shows the lamination log for doing 24 pan - els of product D and 40 panels of product C where each product requires two pressings. At 0.33 hours the first press load entered and at 1.33 hours the first press load left. Table 3 also confirms that 20 panels were used in the first press load. Table 3 shows the results coming out of lamination after the entire run of 64 panels had been completed. Tables 1 and 2 show the state of the lamination queue and the state of the presses where the simulation was stopped at a specific time. These tables are a specific snap shot in time, thus why they are called "dynam- ic database tables." In Table 1, the check marks in the "Send" column show books 6 through 9 making up press load 2. This press load con- sists of two, four-panel books of product C and two, six-panel books for product D. When the panels that are checked are sent to a lamination press, these records are instantly deleted from this table. The accounting of what occurred is recorded and confirmed in the lamination log in Table 3. The information in Table 2 also changes dynamically and can be watched just like a movie or game film that a coach might study, going backwards and forwards, slow and fast, stopping and starting. One more thing to observe in Table 1 and Table 3 are mixed lot and mixed jobs in the press loads just as would oc- cur in your own shop. What happens if you make an oddball com- bination of press cycles? Table 4 shows the lami- nation log results where product C used press cycle C for the first lamination and press cycle B for the second lamination and product D used press cycle D for the first lamination and press cycle C for the second lamination. So books for the second press lamination of product D can be combined with books for the first press lami- nation of product C to make press loads. This is an ugly situation to try to map on paper con- sidering each press cycle has a different time, so different presses will become free in a seem- ingly unpredictable sequence. All of this unpre- dictability, that has nothing to do with random variation or noise, can be predicted within the model and alternative scenarios can be studied as many times as you wish. What impact will more products and press cycles, more or fewer press openings, more or fewer presses, different process and routing se- quences, and all sorts of other possibilities, have on your entire plant? That will be the subject of a future article on modeling and simulation. For now, you may begin to believe that it is possible to model your plant to a high level of detail so that your operations can become more predict- able. And who knows; you might find a pleas- ant, unpredicted surprise that adds plenty to the quality of your life and the quality of your business. PCB